Personalisation on a ‘micro’ level has become fundamental in user-centred design. In fact, around 70% of people expect some level of personalisation now, according to a recent Medium report. Most of us appreciate the gentle reminder that our annual subscriptions are coming up for renewal. For commuters, it’s useful to be able to see recent train trips and rebook at the tap of a button. And automated notifications when we are due for regular health screenings are a valuable part of preventative care. Each of these interactions harnesses personal information in a way that improves the overall experience, helping us to complete tasks with less friction.
When we think of personalisation, we no longer simply think of a signed-in experience that welcomes us by first name. With technology moving at a rapid pace, and access to rich audience data, there’s so much more that can be done. Hyper-personalisation is becoming more common, with real-time responses to behavioural triggers, contextual factors and more. Whether it’s through intelligent alerts, tailored recommendations or dynamic user flows, the scope of personalisation is wide-ranging, and now spans far beyond early adopters like the retail sector.
Over the past decade, Spotify have launched and iterated on their ‘concerts near you’ feature, combining listening data with location data to bring added value to their users through push notifications. Global research software, Qualtrics, now offers real-time sentiment scoring across multiple channels, making it easier to spot frustrated users and route them to senior agents more quickly. Dynamic user journeys have been utilised across both public and private sector domains, pushing users down specific paths based on their needs and behaviours.
But where is the line between personal and intrusive? How can AI be leveraged to make personalisation more scalable? And how can we de-risk personalisation to maintain a positive, compliant experience?
There are a couple of common pitfalls to be aware of:
The intrusivity trap
Research studies have found that when personalisation is ambiguous and unpredictable, it can trigger a more negative response from users. Users experience confusion over how a service provider got their information. From spookily-timed adverts to emails that almost seem to read your mind, there is a fine line to tread between enabling a user to complete a task more easily, and creating distrust.
How can this be avoided?
Having a transparent approach to data collection is critical, not just for compliance, but to remove that feeling of ambiguity. If your audience has a clear understanding of the data you are collecting, they are less likely to feel sceptical. Next, consider the value that personalisation tactics will bring to your audience. Do they solve a known user need, for example, shortening time-to-completion or reducing decision fatigue? Is behavioural data being used in a logical way to bring users information they are interested in? This can be validated through user research prior to implementation.
The mix up
Have you ever had an interaction where personalisation has gone wrong? Perhaps you were recommended something that was time-sensitive, but the recommendation came too late. Or you received a notification that still had placeholder text, where it should have had your information. When inaccurate or irrelevant information is presented, the impact can be huge. Audiences may quickly lose trust in your brand, or worse, you could be at risk of a GDPR breach.
How can this be avoided?
Building a data infrastructure with robust segmentation, governance and hygiene is essential. Without this, it’s difficult to execute effective personalisation and misfires are more likely. Designing with a privacy-first mindset is important to building trust. Having the right technology stack in place can also mitigate risk. Enterprise platforms like Salesforce, for example, heavily invest in data protection frameworks. Having a single source of truth for each user enables better data management, which in turn powers more accurate and reliable personalisation.
How technology is enabling hyper-personalisation
Evolving and emergent technology has taken us from simple, rule-based strategies to more advanced models:
Agentic workflows
Agentic AI can be instrumental in achieving a highly personalised experience, in particular when providing support to service users. By referencing data captured during previous interactions, agentic software can present users with tailored recommendations, helping to resolve issues or tasks more efficiently.
Example: Cleo, the financial management app, uses non-linear agentic architecture to work through complex tasks alongside proactive reasoning to make highly personalised recommendations.
Predictive analytics
Predictive analytics is powering more strategic service delivery for many organisations. Rather than have each user move through a uniform flow, data trends can be used to anticipate where a specific user should go next and what touchpoints they should interact with.
Example: The Debt Management Office recently announced a new strategy that will use predictive technology to identify where people may be at risk of falling into debt, and respond by providing more tailored support and resources.
Generative AI
Personalisation becomes more scalable with the use of generative AI, where nuanced multi-variant content can be created and served to users more quickly. For example, previously browsed content themes can be mapped onto content to create a bespoke interaction that nudges the user along.
Example: PODS, the moving and storage company, uses generative AI to power its Smart Billboard, dynamically populating ad content tailored to whichever neighbourhood it passes through.
What next?
Personalisation should feel like a seamless part of the user experience. To achieve this while avoiding some of the pitfalls covered above, it’s important to build the right foundations:
- Interrogate your data and audit how well it is captured, segmented and maintained.
- Map your current journeys to identify steps where personalisation may be beneficial.
- Test your hypotheses with real users to understand their expectations, thoughts and feelings.
At Hippo, our service design, data and technology teams help organisations understand where personalisation can add value, where it may introduce risk and how to design experiences that feel clear, trusted and genuinely helpful for users.
